the new third generation: post-1965 immigration and the next … · 2017-12-19 · assimilation...

The New Third Generation: Post-1965Immigration and the Next Chapter inthe Long Story of Assimilation

Tom!as R. Jim!enezStanford University

Julie ParkUniversity of Maryland at College Park

Juan PedrozaStanford University

Now is the time for social scientists to focus an analytical lens on thenew third generation to see what their experiences reveal about post-1965 assimilation. This paper is a first step. We compare the house-hold characteristics of post-1965, second-generation Latino andAsian children in 1980 to a “new third generation” in 2010. Today’snew third generation is growing up in households headed by parentswho have higher socioeconomic attainment; that are more likely to beheaded by intermarried parents; that are less likely to containextended family; and that, when living with intermarried parents, aremore likely to have children identified with a Hispanic or Asian labelcompared to second-generation children growing in 1980. We usethese findings to inform a larger research agenda for studying the newthird generation.

For the better part of the last two decades, scholars have debated whetherthe post-1965 immigrant groups, which come mostly from Latin Ameri-can, Asia, and the Caribbean, are following the assimilation path blazedby earlier waves of immigrants. Some maintain that today’s immigrantgroups are following multiple assimilation trajectories, with large numberswinding up permanently poor and negatively racialized (Portes and Zhou1993; Rumbaut 2005; Haller, Portes, and Lynch 2011). Others are moresanguine, noting the upward mobility from the first to second generation,resiliency, and economic outcomes hardly align with dire predictions of

© 2017 by the Center for Migration Studies of New York. All rights reserved.DOI: 10.1111/imre.12343

IMR Volume ** Number ** (Fall 2017):1–40 1

some among the post-1965 second generation ending up in a “rainbowunderclass” (Kasinitz et al. 2008; Alba, Kasinitz, and Waters 2011;National Academies of Science, Engineering, and Medicine 2015).

Yet the disagreement over assimilation in the post-1965 immigrationera is over something that is yet to happen. As historical assimilation ofthe Southern and Eastern European immigrant groups of the early part ofthe twentieth century showed, assimilation is a multigenerational processthat continues to unfold beyond the second generation. But the debaterages on because scholars have had to study the post-1965 assimilationusing only data on first and second generations — until now. It has beennearly a generation since the first statements on the assimilation of the“new second generation” of the post-1965 immigrants (Gans 1992;Portes and Zhou 1993) kicked off these debates. And now, US society iswitnessing the rise of the grandchildren of the post-1965 immigrants —the “new third generation.” With the tremendous ethnoracial and classdiversity of the post-1965 immigrants, the new third generation will writethe next and most significant chapter of contemporary assimilation.

Now is the time for social scientists to focus an analytical lens onthe new third generation to see what their experiences reveal about post-1965 assimilation. We offer a first step in that direction through empiricalanalysis and a larger research agenda for studying this new third genera-tion. Our empirical analysis draws on 1980 decennial US Census data toexamine the household characteristics of the second generation. We com-pare those characteristics to data on third-generation children in the2008–2013 (“2010” hereafter) Current Population Survey (CPS). We usethese data in descriptive and multivariate analyses, focusing on changes incharacteristics of the households in which the second generation grew upin 1980 compared to a third generation in 2010. We also examine differ-ences in the way intermarried parents ethnoracially identify their childrenon survey forms between these two periods. We find that the new thirdgeneration of the post-1965 immigration era is, on virtually every mea-sure, growing up in better household circumstances than the second gen-eration 30 years earlier. The new third generation is growing up inhouseholds with parents who have higher socioeconomic attainments thanthe second-generation children did three decades earlier. They are alsomore likely to be living with intermarried parents, and they are less likelyto be living in extended-family households. These findings indicate a pathtoward greater assimilation for the new third generation. But when itcomes to trends in ethnoracial identification, the relevance of prevailing

2 INTERNATIONAL MIGRATION REVIEW

assimilation theories is less clear. Third-generation children living inhouseholds with one non-Hispanic and one Hispanic parent in 2010 aremore likely to be labeled as Hispanic on survey forms than theirsecond-generation counterparts three decades earlier. Households with oneAsian and one non-Asian parent in 2010 show similar patterns: third-gen-eration children living with intermarried parents in 2010 more often haveassigned to them an Asian label than second-generation children livingwith intermarried parents in 1980. These latter findings point assimilatorypaths for the new third generation that may both converge and divergefrom patterns observed among the European-origin third generation.While instructive, these findings and their implications come from ouranalytical glimpse into the lives of the new third generation. We thus layout a research agenda for a new era of assimilation research that addressesthe empirical and theoretical imperatives for fuller understanding of post-1965 immigrant groups.

IMMIGRATION THEN AND NOW; ASSIMILATION THENAND NOW

The touchstone for studying the post-1965 immigrant assimilation is thelarge-scale immigration from Europe that took place mostly between 1880and 1920 (Perlmann and Waldinger 1997; Perlmann 2005). That histori-cal immigration wave was the basis for the development of an assimilationtheory that endured for much of the twentieth century (Park and Burgess1969 [1921]). The first large-scale empirical study of assimilation that fol-lowed was Warner and Srole’s (1945) study of “Yankee City,” whichincluded data on the intergenerational change of the city’s immigrantgroups. The second-generation children and some third-generation grand-children of the European immigrants made it into the study, leading War-ner and Srole to assert that assimilation proceeded in step fashion, witheach generation passing along its advantages and disadvantages to thenext, and shaping the pace of assimilation toward a common destination.

Subsequent studies of assimilation clarified the importance of genera-tion-since-immigration in the assimilation process. Writing as the third-generation descendants of the European immigration were coming of age,Milton Gordon (1964) identified a series of interrelated stages of assimila-tion, of which “structural assimilation” — “large-scale entrance into cli-ques, clubs, and institutions of the host society, on [a] primary grouplevel” — and “identificational assimilation” — “[d]evelopment of a sense

THE NEW THIRD GENERATION 3

of peoplehood based exclusively on [the] host society” — were para-mount. The full solidification of assimilation theory happened two dec-ades later, when scholars took a close look at the role that ethnoracialorigin played in the lives of adult third- and fourth-generation descen-dants of the Southern and Eastern European immigration. Those studiesoverwhelmingly showed that ethnoracial origins had entered a “twilight”(Alba 1985), forming a symbolic and optional part of identity that scar-cely had negative effects on life chances and opportunities (Gans 1979;Alba 1990; Waters 1990).

More recent scholarship has turned attention to historical immigrantpopulations that never made it into the canon of classic assimilation the-ory. Among the descendants of historical waves of Chinese and Japaneseimmigrants, there is overwhelming evidence of economic progress acrossgenerations (Alba and Nee 2003), even though some of these individualsare still regarded as “forever foreigners” (Tuan 1998). Third- and later-generation descendants of historical waves of Mexican immigrants displayintergenerational progress, along a host of measures (Macias 2006; Dun-can and Trejo 2007; Telles and Ortiz 2008; Lichter, Carmalt, and Qian2011; Qian and Lichter 2011). Yet even with this intergenerational pro-gress, Mexicans have not reached parity in their economic outcomes withUS-born whites. The incompleteness of that assimilation is almost cer-tainly due to historic discrimination. It may also be partly due to the waythat the deleterious effect of unauthorized status on the socioeconomicattainments of earlier generations lingers on in subsequent generations(Bean, Brown, and Bachmeier 2015).

Just as the final stages of the Southern and Eastern European assimi-lation came into clearer focus, a new era of assimilation theorizing wasbeginning — this one for the post-1965 immigration. Named for the yearin which liberalizing federal immigration reforms were passed,1 the post-1965 immigrants came overwhelmingly from non-European origins: LatinAmerica (largely Mexico), Asia, and the Caribbean. They also come froma spectrum of class origins, filling up the ranks of the poorest and leasteducated, as well as the wealthiest and most educated people in the Uni-ted States (Portes and Rumbaut 2014). These two features — nonwhite-ness and class diversity — of the post-1965 immigration have led some

1The 1965 Immigration and Nationality Act, also known as the Hart-Celler Act, liftednational-origins quotas passed in 1924, created a preference system for immigrant admis-sions, and established equitable hemisphere and country visa limits.


scholars to revive and adjust the central tenets of assimilation theory tomake sense of the experiences of the second generation. The resulting“segmented assimilation” perspective posits that today’s immigrant groupsassimilate into one of multiple ethnoracial and class-defined segments ofUS society (Portes and Zhou 1993). The class origins of the immigrants,their social and legal reception in the receiving society, and the cohesive-ness of the co-ethnic community together determine segments of societyto which groups assimilate (Portes and Rumbaut 2001). Particularly atrisk of assimilating on a “downward” path into the most precarious ethno-racial and class segments are second-generation Mexicans, some of theCaribbean-origin second generation, and some Southeast Asian groups.Their high poverty rates, negative legal and social reception, and co-eth-noracial communities lacking in social capital make their assimilationespecially precarious, according to the perspective (Portes and Zhou 1993;Portes and Rumbaut 2001).

These claims have their detractors. Among them are scholars whosee the post-1965 second generation as having more in common with thesecond-generation children of Southern and Eastern European immi-grants, in terms of their social and economic trajectories, than the seg-mented assimilation perspective would suggest (Alba and Nee 2003;Perlmann 2005). Others note that the direst assessments of the secondgeneration do not bear out in empirical analyses (Waters et al. 2010).Some even argue that the second generation individuals have an advan-tage. Institutional arrangements with roots in the Civil Rights Movementpositively recognize ethnoracial difference, offering a leg up to the non-white second generation (Waters 2010). And the second generation maypossess a degree of cultural and social flexibility that pays off in multipledimensions of life (Kasinitz et al. 2008; Alba, Kasinitz, and Waters 2011).

Still others argue that the assimilation/nonassimilation dichotomythat characterized research on the second generation from the past twodecades paints on overly simplistic portrayal of the assimilation process.In studying the Mexican-origin population in Los Angeles, Bean, Brown,and Bachmeier (2015) argue that the poor socioeconomic origins, and,importantly, the unauthorized status of some immigrants, delay integra-tion across generations such that it may take more than three generationsfor full integration to occur. Bean, Brown, and Bachmeier (2015) alsoargue that the various social, political, and economic dimensions of assim-ilation do not necessarily operate in lockstep. With the increased institu-tional recognition of myriad individual identities, ethnoracial identity may


not necessarily decline across generations, even as socioeconomic advance-ment attains.

These insights about the assimilation of the post-1965 immigrantsfrom a vast body of research are more like detailed assimilation roadmarkers than definitive destinations. Debates about the assimilation of thepost-1965 assimilation rely almost exclusively on studies of the first andsecond generation. Assimilation is a multigenerational process, a principlethat has receded too far into the backdrop of debates about the fate ofpost-1965 immigration wave. Although comparisons between today’simmigrants and the last great wave that came from Europe are debatable(Foner 2005; Fox and Guglielmo 2012), a central insight gleaned fromthe study of those earlier immigrants is relevant to the post-1965 immi-grants: it may take well more than two generations for the full extent ofassimilation to unfold. Competing views of post-1965 assimilation —from the dourer and the more sanguine — cannot offer a complete pic-ture of post-1965 assimilation because only two immigrant generationshave come of age during the past half-century. Much as Perlmann andWaldinger (1997) noted nearly two decades ago, we argue that the fullextent and character of post-1965 immigration assimilation will emergewith time, as the third and even fourth generation come of age. WhilePerlmann and Waldinger had to rely on prediction, we now have the datato begin assessing the nature of the post-1965 assimilation over three gen-erations. Now is the time for social science research to turn its attentionto the new third generation with the same vigor that has been behindstudies of the second-generation.

STUDYING THE NEW THIRD GENERATION

We embark on that endeavor by examining several important outcomesmeasured at the household level. We cannot distinguish the new third gen-eration of the post-1965 immigration wave from the fourth and later gen-eration from earlier immigration waves unless the new third generation arechildren who coreside with parents. Therefore, second- and third-genera-tion children living with parents serve as the unit of analysis, and weobserve the characteristics of household in which they live: socioeconomicattainment of parents; having intermarried parents; living with extendedfamily; and ethnoracial identification assigned by parents.

Socioeconomic attainment sits at the center of assimilation theories.Classical assimilation theory posits that each successive immigrant


generation will improve its socioeconomic attainment, indicated by moreeducation and greater income. The segmented assimilation perspective, onthe other hand, might predict socioeconomic stagnation or even declineinto the third generation for Latinos, and a racialized form of identity inresponse entrenched racism. Asians with elevated socioeconomic origins,according to segmented assimilation, would find a place in the middleclass, while perhaps maintaining a connection to ethnoracial identity thatresembles that of the immigrant generation (Portes and Rumbaut 2001).

Another indicator of assimilation is whether the household contextinfluences ties to the immigrant ethnoracial community through the pres-ence of grandparents or other relatives in the household. The presence ofextended family can be a strategy to cope with economic hardship or away to get ahead by pooling resources. But the presence of extended fam-ily can also help maintain connections to ethnoracial culture because itmay increase the need to speak a language other than English at home aswell as observe other cultural or social ties to the ethnoracial community(Zhou and Xiong 2005).2 Straight-line assimilation theory would predicta decline across generations in the presence of extended family membersin the household. Segmented assimilation, which views sustained attach-ment to a co-ethnic community as a benefit to socioeconomic attainment,might predict a sustained presence of extended family among upwardlymobile groups, while a decline in the presence of extended family amonggroups experiencing downward assimilation.

Another key dimension of assimilation is intermarriage. Gordon(1964) included marital assimilation as one of the final stages of theassimilation process. According to classic assimilation theory, intermarriageshould increase with each generation. The segmented assimilation perspec-tive does not offer a clear predication on intermarriage. But the majortenets of the perspective suggest that groups experiencing downwardmobility will have very low levels of intermarriage because of their socialisolation; groups experiencing upward mobility might see an accelerationof intermarriage after the second generation. Studies of intermarriage ratesin the post-1965 era offer a glimpse of these patterns, showing higher

2In supplementary analyses not reported here, we find only small differences in unemploy-ment rates of the head of household (one percentage point in 1980; two percentage pointsin 2010), and small differences in the proportion of household below the poverty line(two percentage points in 1980; one percentage point in 2010) with extended familymembers and those without. These analyses are available upon request.


rates of out-marriage for the US-born than those for immigrants amongHispanics and Asians (Qian and Lichter 2007). This rise was particularlyamong third-plus-generation Hispanics whose immigrant ancestors werefrom a previous immigration wave; intermarriage among the second gen-eration grew more slowly during the 1990s because the rapid growth ofAsian and Hispanic immigrant populations provided greater opportunityfor the second generation to find marriage partners of the same ethnora-cial origin (Qian and Lichter 2011).

Intermarriage is relevant for the kinds of identity that the childrenof intermarried couples take on (Lieberson and Waters 1988; Telles andSue 2009).3 Ethnoracial boundaries can reconfigure depending on howthe children of intermarriage think of themselves in ethnoracial terms (Leeand Bean 2010). Of course, just as intermarriage is intimately linked toother assimilation processes, so too might differences in ethnoracial identi-fication among the children of intermarriages be linked to differences insocioeconomic attainment (Duncan and Trejo 2011; Emeka and Vallejo2011). Classic assimilation would predict a weakening of ties to the eth-noracial identification of the immigration generation into the third gener-ation. Segmented assimilation would predict that Latinos maintain astrong attachment to a Latino ethnoracial identification in response topersistent discrimination. Asians may also maintain some attachmentbecause, according to segmented assimilation, attachment to the ethnora-cial identification of the immigrant generation is tied to positive socioeco-nomic outcomes.

Analytical Approach

Due to data limitations, countless studies have attempted to measureassimilation across generations using cross-sectional data, often only mak-ing the distinction between immigrants and the US-born of a particularethnoracial group (LaLonde and Topel 1992; Kao and Tienda 1995;Garcia 2011). The cross-sectional approach approximates rates for eachimmigrant generation at the same point in time, but it does not effectivelycapture true intergenerational change. The cross-sectional approach mayalso underestimate intergenerational assimilation, compared to an

3Cohabitation, which has been on the rise in the last two decades, is also an importantindicator of the durability of ethnoracial boundaries. Still, the legal, contractual nature ofmarriage makes it a more compelling indicator of those boundaries.


immigrant-generation-cohort approach (Parrado and Morgan 2008; Parkand Myers 2010), which more accurately captures difference betweencohorts of parents and children by also accounting for age. Furthermore,a cross section of immigrant generations at a single point in time wouldmost likely include a third generation from a previous wave of immigra-tion. To determine whether straight-line assimilation captures the experi-ence of the post-1965 immigration era, we utilize the immigrant-generation-cohort method (Park and Myers 2010), which simultaneouslymeasures the intergenerational mobility between immigrant generations atcomparable ages, relative to a reference group in the same time period.The immigrant-generation-cohort method effectively makes, in this case,comparisons between a young third generation and a young second gener-ation 30 years earlier (Park and Myers 2010).

Data limitations pose a challenge to studying the new third genera-tion in terms of parents’ socioeconomic attainment, the presence ofextended family, intermarriage, and ethnoracial identification of children.Perhaps the most significant challenge is that a unique third generation isnot readily identifiable in most publicly available survey data. The mostprevalent way of identifying the third generation in large surveys, such asCPS, is to use questions about respondents’ place of birth in combinationwith their parents’ place of birth to identify a “third-plus generation,”which includes individuals who were born in the United States to US-born parents. However, this third-plus generation encompasses individualswhose grandparents are foreign-born as well as individuals whose ancestralroots in the United States extend back more than three generations. It isthus virtually impossible to identify a “pure” adult third generation ofindividuals whose grandparents were first-generation immigrants.4

The CPS, a government survey administered jointly by the USBureau of Labor Statistics and the US Census Bureau, began asking aboutparents’ place of birth in 1994. But it is still impossible to identify thepure adult third generation from these data. It is, however, possible toidentify third-generation children residing in households headed by atleast one-second-generation parent through the use of the parental nativityquestion asked of both the children and parents in the household. We usethis approach in our analysis. Due to the small sample size of the CPS,especially in its ability to capture adequate numbers of a unique third

4This task is made even more difficult by the fact that the US Decennial Census stoppedasking about parental nativity after the 1970 decennial census.


generation, we pool six years of data from 2008 to 2013, which we referto as “2010” in the analysis.5 Our new third-generation sample includesall US-born children who are co-residing with their second-generationparents.6 We include parents between the ages of 25 and 54 because theseare the prime working ages and also the ages at which parents are mostlikely to have children living in the household.

To determine generational differences from the second to the thirdgeneration, we compare the family and household context of third-genera-tion children in 2010 to those of second-generation children in 1980.Drawing on 1980 decennial Census data, our second-generation sampleincludes all US-born children who are coresiding with their immigrant par-ents. Our analysis thus captures the household circumstances in which thenew third generation is coming of age today compared to the householdcircumstances in which a second generation came of age 30 years earlier. Itis important to consider what these data capture more precisely. Ourapproach does not allow for the direct observation of kinship and thus doesnot directly measure change from parents to their children (Duncan 1966).Still, our methodological approach of observing successive immigrant gen-erations across time samples from each generation cohort broadly capturesintergenerational change (Smith 2003, 2006a; Park and Myers 2010).While a strength of the Census and CPS, respectively, is that they allow usto identify households in which a unique second and third generationreside, these data sources do not provide detailed information on the cir-cumstances of the individual children in these households. Indeed, the dataon household characteristics that we analyze significantly reflect the behav-iors (education, income, intermarriage, ethnoracial identity) of the parentsof our second- and third-generation populations of interest. But the charac-teristics of household in which children reside are highly predictive of adultoutcomes (Chetty et al. 2014a), and so the behavior of these parents, as

5We use data from the Minnesota Population Center’s Integrated Public Use MicrodataSeries (IPUMS). To capture second-generation households in the post-1965 immigrationera, we use the 1980 PUMS 5 percent sample (Ruggles et al. 2015) and the IPUMS CPS(March Supplement; King et al. 2010) from 2008 through 2013 (sample is limited to per-sons who are in their first four months of the survey to avoid replicating cases in consecu-tive years).6If children are residing with mixed-generation parents, they are assigned according to theparent with the higher generation status. For example, a child with one immigrant parentand one-second-generation parent is defined as being a part of the third generation. Sup-port for this rationale can be found in Ramakrishnan (2004).


reflected in the indicators we use, is also telling of the future trajectory ofthese children as they enter adulthood.

We use the immigrant-generation-cohort method developed by Parkand Myers (2010) to compare the kinds of households in which the newthird generation is reared to those of their parents’ cohort a generationearlier. To place the multivariate findings for immigrant generations incontext, we compare Hispanic and non-Hispanic Asian (“Asian” hereafter)immigrant generations to third-plus-generation non-Hispanic whites (or“white” in the rest of the paper) as the reference group. This referencegroup acts as a proxy for the societal standard of the historical time per-iod. As the sample of this reference group is much larger than the sampleselected for immigrant generations, we randomly selected 5 percent of the1980 sample and 20 percent of the 2010 sample for statistical analyses.Our unit of analysis is children in these specified households. It is impor-tant to note that our analysis does not attempt to capture the characteris-tics of all immigrant adults in 1980, or second-generation adults in 2010.Instead, our data capture households in which there are immigrant parentsof the second generation in 1980, and second-generation parents of thethird generation in 2010. Our analytical focus is on the second- andthird-generation children and the characteristics of the households inwhich are growing up.

Beyond socioeconomic indicators of assimilation, which are welldocumented with this methodology (Park and Myers 2010; Park, Myers,and Jim!enez 2014; Park, Nawyn, and Benetsky 2015), the coresidence ofthe young third generation with their second-generation parents providesan ideal opportunity to examine marriage and household structure asforms of assimilation of Hispanics and Asians. Therefore, we look at thepresence of extended family members coresiding with third-generationchildren in 2010 and compare them to the shares for the second-genera-tion children a generation earlier (in 1980). We also examine the share ofthird-generation children coresiding with intermarried parents. Finally, weexamine the ethnoracial labels that parents assign to children in thesehouseholds. In contrast to what the Census sets forth, we treat Hispanics,regardless of their ethnoracial sub-identification, as a single “race” groupsuch that anyone who answers affirmatively to the Hispanic question isassigned to a Hispanic category, regardless of the race category selectedfrom the separate race question. Likewise, we treat anyone who identifiedwith any of the race categories conventionally subsumed under the pan-Asian category as broadly “Asian.”


For our three outcomes, we use the following logistic regressionmodel:

ðOÞ ¼ Year þ Generation þ Year % Generation þ Age

þMother ’s Education þ Region

Our independent variables of interest are temporal. Year is theobservation period (1980 is the reference year and 2010 = 1) which cap-tures the period effects. Generation represents the generation-since-immi-gration status and the referent non-Hispanic whites in both 1980 and2010 (non-Hispanic whites are the reference group and the parents of thesecond generation in 1980 and the parents of the third generation in2010 = 1). The interaction term, Year 9 Generation, is the net effectchange for these generations beyond the change observed for the whitereference group. Age is included to control for any age distribution effectson the outcome. Mother’s education is a proxy for socioeconomic status.Region is a control for any geographic differences.

We first explore the presence of immigrant grandparents or otherextended family members in these households. We compare third-genera-tion children in 2010 to second-generation children in 1980, modelingthe determinants of living with extended family. We include both one-and two-parent households for these analyses and determine whethereither is more likely to live with extended family. We then examine howsocioeconomic status (proxied by mother’s education or the education ofthe head of household when one-parent households are included in thesample) is associated with each of the three outcomes. Given that theremay be significant differences for these outcomes across the United States,we include Census regions.

A second set of analyses examines the determinants of living withintermarried parents. Intermarriage is defined here as marriage betweenindividuals who do not identify themselves of the same race — in the caseof Hispanics, any marriage between someone identifying with any His-panic category and someone identifying with no Hispanic category is anintermarriage; in the case of Asians, any marriage between someone iden-tifying with any Asian category and someone who identifies with no Asiancategory at all is an intermarriage. We do not count as intermarriagesunions among different subcategories of Asians and Hispanics (e.g.,between Puerto Ricans and Mexicans; between Japanese and Chinese).Therefore, our estimate of intermarriage should be regarded as


conservative relative to the scale of intermarriage had we counted anycross-category marital union as intermarriage.

The final set of analyses explores the ethnoracial labels assigned tochildren when their parents come from different ethnoracial ancestries —when one parent is Hispanic and the other parent is not Hispanic; whenone parent is Asian and one parent is not Asian. We examine these ethno-racial labels for second-generation children in 1980 and third-generationchildren in 2010. We model the odds of third-generation children receiv-ing a Hispanic or Asian label in 2010 relative to 1980. We also look atthe household-level factors predicting those labels. It is important to notethat the 2010 CPS offered more choice in selecting ethnoracial labels thanthe 1980 census. Beginning in 2000, the census and the CPS permittedrespondents to check multiple racial categories. Thus, intermarried parentsin 1980 were forced to select a single racial category for their child, butparents 2010 could select multiple racial categories. The differencebetween the categories available in the 1980 census and the ability toselect multiple racial categories in government surveys after 2000 —including in the CPS — is not relevant for households with a Hispanicparent because the census and CPS treat Hispanicity and race as distinc-tive, where parents select whether their children are Hispanic and theirchild’s race in response to separate questions. The difference in availableracial categories is relevant to households with Asian parents because thecensus and CPS treat the Asian subcategories as distinctive racial cate-gories, allowing parents to select an Asian category in combination withother racial categories (e.g., Chinese and white; Japanese and black). Thedifference in the availability of categories for these households meant thatwe had to make decisions about how to regard children living in Asianintermarried households who were assigned an Asian category in combina-tion with other racial categories. Following assimilation theories, we wereinterested in the degree to which there is some maintenance of an identifi-cation with the Hispanic or Asian category over time and across genera-tions. Our analysis thus treats children who were identified as Asian incombination with another racial category in 2010 the same as childrenwho were given an Asian label exclusively. Because parents who selectAsian in combination with other categories for their children are not dis-regarding an Asian ancestry, we count these children as among those whoappear to maintain an identification with Asian ancestry.


Overview of the New Third Generation

Table 1 includes the household characteristics of second-generation chil-dren in 1980 and third-generation children in 2010 compared to thewhite reference group in each time period. Households include both one-and two-parent households. The vast majority of the Asian and Latinonew third generation under age 19 are young children: two-thirds areunder the age of 10. And because data limitations make it impossible toidentify post-1965 third-generation adults separate from individuals whoare more than three generations removed from the immigrant generation,

TABLE 1HOUSEHOLD CHARACTERISTICS FOR THE SECOND GENERATION IN 1980 AND FOR THE THIRD GENERATION

IN 2010 (INCLUDES BOTH ONE- AND TWO-PARENT HOUSEHOLDS) BOTH ONE- AND TWO-PARENT HOUSE-

HOLDS

Hispanichouseholds

Asian, non-Hispanichouseholds

White, non-Hispanichouseholds

2ndGeneration(1980)

3rdGeneration(2010)

2ndGeneration(1980)

3rdGeneration(2010)

3rdGeneration(1980)

4th+Generation(2010)

Two-parenthousehold (%)

75 73 93 88 88 80

Age (mean) 7 7 6 7 9 8Under age 10 (%) 65 65 76 69 54 56Head of householdage (mean)

37 36 37 39 37 39

Family size (mean) 5.2 4.5 4.7 4.4 4.6 4.3Number of children(mean)

3.2 2.6 2.4 2.3 2.6 2.5

Extended family (%) 16.3 9.9 20.6 11.0 4.3 4.1Household head withBA+ (%)

6.4 20.2 49.5 55 22.8 39.5

Unemployed head ofhousehold (%)

7.2 7.4 2.3 4.6 3.9 5.9

Working adults perhousehold (mean)

1.2 1.4 1.6 1.6 1.4 1.5

Household income(median, 2010dollars, thousands)

$40 $49 $75 $94 $66 $77

Poverty status(% belowpoverty line)

32.1 21.9 8.0 7.2 9.5 10.4

Observations(unweighted)

66,124 6,645 18,425 1,754 1,843,364 82,871

Source: Authors’ calculations of March Current Population Survey (2008–2013) and 1980 decennial Census. Fig-ures by generation status include one- and two-parent households with children aged 16 and under. Figures are lim-ited to households where head of household is between 25 and 54 years old.


we observe the family and socioeconomic characteristics of the householdsin which they grow up.

When comparing the household characteristics of Hispanics andAsians to non-Hispanic whites of a later generation, household structurediffers across the groups, and some measures change over time. The meanage of heads of household is in the late thirties. Family size decreases andconverges over time for both Asian and Hispanic groups (to 4.4 and 4.5in 2010, respectively). Both groups have approximately two children perhousehold, and, notably, there is a decrease (from 3.2 to 2.6 children)among Hispanics from the second to the third generation. The averagenumber of children for Asians and whites changes little, and the threegroups converge to an average of between 2.3 and 2.5 children per house-hold by 2010.

The share of married-couple households declined across all groups,although whites experienced the greatest drop. It is important to note herethat we applied age (for children and parents) and generation-status filtersthat resulted in an analytical sample with lower rates of single-parent-headed households than exists in a full sample that does not havethese filters. However, the downward trend of two-parent households inour sample conforms to a more pronounced trend in a sample withoutselection filters.7 Extended-family households are more common amongHispanics and Asians than among whites. Very few white households livewith extended family members (4.1% in 2010), while Hispanics andAsians are much more likely to do so. For example, 11 percent of Asianhouseholds with third-generation children live with extended family mem-bers (compared to 9.9% for Hispanic households with third-generationchildren). These shares of extended-family households decreased from20.6 percent among Asian households with second-generation children in1980 (higher than the 16.3% for Hispanic households with second-generation children).

Hispanics and Asians share some similar household characteristics(e.g., family size, number of children, extended-family households) towhites. However, immigration policy and historical relationships between

7We recognize that the share of married-couple households is higher for our sample thanwhat is conventionally reported for race groups in the United States (The Annie E. CaseyFoundation’s Kids Count data; see http://datacenter.kidscount.org/data#USA/1/2/3,6,5,4/char/0). This is mainly due to the age and generation-status filters we use to restrict oursample. Our full sample without these restrictions results in married-couple householdrates that are more consistent with others who report this.


sending countries and the United States has produced a socioeconomicprofile for Hispanic immigrant generations that differs substantially fromthat of Asian immigrant generations in the post-1965 immigration era.Table 1 also includes levels of parental education across time and genera-tions. Hispanic parents of the second generation in 1980 had low educa-tional attainment, but there was substantial intergenerational mobility by2010. Almost half of Asian householders had at least a college degree in1980. Asian immigrant generations also experienced intergenerationalmobility, as did white householders in the same time period. The differ-ences in human capital across these groups help to explain the differencesin labor force and economic characteristics of these households, includingsimilarly low yet persisting poverty rates for both whites and Asians.

Living with Extended Family

While the descriptive statistics in Table 1 are suggestive, we offer multi-variate analyses to more precisely capture differences across time and gen-eration. We begin with the predictors of living with extended family.Table 2 includes the odds ratios of living in extended-family householdsfor Hispanic children relative to white children.

As shown in Table 1, very few whites live in extended-family house-holds, and the odds of children living in such households has decreasednominally over time (0.9 times the odds of 1980 in 2010). Relative tochildren in white households, children in Hispanic households have 4.27times the odds of living with extended family members. All third-genera-tion children in 2010 have 0.59 times the odds of living with extendedfamily relative to second-generation children in 1980. The associationbetween socioeconomic status and living with extended family members ismixed with differences for educational attainment and economic status.Model 4 shows that having a householder with a bachelor’s degree reducesthe odds of living with extended family by 43 percent. On the otherhand, those who live in households below poverty are 80 percent lesslikely to live with extended family (but the effect of poverty status is onlysignificant after accounting for mother’s education, as Models 3 and 4 inTable 2 show). Therefore, it is not necessarily those who have the highesteducational or low economic status who are living with extended family.Rather, it may be those who do not have high educational attainment,but do have some economic stability, who coreside with extended familymembers. Lastly, there is a statistically significant difference between those


TABLE2

ODDSRATIO

SOFLIV

ING

INANEXTENDED-F

AMILYH

OUSE

HOLDCHILDREN

WIT

HATLEAST

ONEH

ISPANIC

PARENT

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Year(1

=20

10)

0.959

0.96

01.07

81.068

0.892

0.891

1.006

0.99

3Generation(1

=Hispanic

3rdgenerationin

2010

;Hispanic2n

dgeneration

in19

80)

4.270***

4.34

6***

4.04

4***

3.703***

4.901***

4.963***

4.571***

4.18

0***

Year9

generation

0.591***

0.58

6***

0.56

2***

0.562***

0.487***

0.484***

0.469***

0.47

1***

Age

1.012

1.01

21.00

71.008

1.017*

1.017*

1.013

1.01

3Poverty

status

(1=below

povertylin

e)0.92

40.79

5*0.802*

0.914

0.798

0.80

0

Householder

education

(1=bachelor’sdegree)

0.43

8***

0.437***

0.474***

0.47

4***

One-parent

household

2.268***

2.746***

2.466***

2.43

9***

One-parent

household

9year

1.074

1.079

1.024

1.04

7

One-parent

household

9generation

0.465***

0.548***

0.585***

0.60

1***

One-parent

household

9year

9generation

2.205***

1.953***

1.932***

1.86

2***

One-parent

household9

age

0.970*

0.966*

0.970*

0.97

0*One-parent

household

9povertystatus

0.630*

0.663

0.66

9

One-parent

household

9householdereducation

1.011

1.01

0

Region(ref:Northeast)

Midwest

0.733**

0.73

9*South

1.009

1.02

9West

1.140

1.14

9Con

stant

0.042***

0.042***

0.051***

0.054***

0.037***

0.037***

0.044***

0.046***

Observation

s18

1,51

118

1,51

118

1,51

118

1,51

118

1,51

118

1,51

118

1,51

118

1,51

1Log

likelihood

&7.210e+0

6&7.21

0e+0

6&7.11

9e+0

6&7.102e+0

6&7.133e+0

6&7.113e+0

6&7.042e+0

6&7.02

5e+0

6

Source:Authors’calculations

ofMarch

Current

PopulationSurvey

(2008–

2013)and1980

decenn

ialCensus.Notes:***p

<0.001,

**p<0.01,*p

<0.05.


living in the Midwest and those to the Northeast (the Midwest has 0.73times the odds of the Northeast).

Models 5 through 8 examine how one-parent households may differin terms of their propensity to live with extended family members. One-parent families have more than twice the odds of living with extendedfamily relative to two-parent families. This is less the case for Hispanichouseholds in 1980, but dramatically increased by 2010. Socioeconomicstatus (household poverty, parent education) in one-parent householdsdoes not function in significantly different ways than it does for two-parent households. It is striking to note that regardless of socioeconomicstatus and parental structure, those in Hispanic families are consistentlymore than four times as likely to live in extended-family households thanwhite families.

Table 3 presents the same outcome for children living in Asianhouseholds. Children in these households have greater odds of living inmultigenerational or extended-family households than other ethnoracialgroups.

Model 1 confirms that all children in Asian households have 5.56times greater odds of living with extended family than whites. Third-generation children are half as likely to live with extended family as sec-ond-generation children in 1980. Consistent with findings for Hispanichouseholds, children in Asian households with householders who have abachelor’s degree have odds of living with extended family that are 0.38those of children living with less educated householders (Model 4). Theassociation between poverty status and the likelihood of living withextended family is not statistically significant across models for Asians,although the odds ratios are consistent with results for Hispanics.Unlike the findings for Hispanic households, the introduction ofsocioeconomic status increases the odds of Asian families living inextended-family households to over seven times. The higher educationalattainment of Asian households suppresses the degree to which Asianfamilies live in extended-family households. Once controlling for educa-tional attainment, the odds of living in extended-family householdsincreases from 5.56 to 7.02. As observed among Hispanics, living out-side the Northeast and in the Midwest yields just 0.70 times the oddsof living with extended family among Asians compared to living the inthe Northeast. Also consistent with outcomes for Latinos, children inAsian one-parent households are more likely to live with extended


TABLE3

ODDSRATIO

SOFLIV

ING

INANEXTENDED-F

AMILYH

OUSE

HOLDCHILDREN

WIT

HATLEAST

ONEASIANPARENT

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Year(1

=20

10)

0.95

40.952

1.084

1.068

0.885

0.88

71.024

1.002

Generation(Asians)

(1=Asian

3rd

generation

in2010;

Asian

2nd

generation

in1980)

5.56

2***

5.581***

7.024***

6.300***

6.026***

6.02

9***

7.576***

6.793***

Year9

generation

0.50

0***

0.501***

0.456***

0.455***

0.466***

0.46

6***

0.424***

0.424***

Age

0.99

60.996

0.988

0.988

0.994

0.99

40.985

0.985

Poverty

status

(1=below

povertylin

e)1.176

0.899

0.912

1.09

80.844

0.853

Householder

education

(1=bachelor’sdegree)

0.381***

0.383***

0.395***

0.399***

One-parenthousehold

1.943***

2.261***

1.878**

1.846**

One-parenthousehold

9year

1.077

1.07

61.001

1.033

One-parenthousehold

9generation

0.562***

0.56

1***

0.501***

0.490***

One-parenthousehold

9year

9generation

2.352*

2.25

5*2.319*

2.306*

One-parenthousehold9

age

0.988

0.98

40.995

0.995

One-parenthousehold

9povertystatus

0.64

90.745

0.745

One-parenthousehold

9householdereducation

1.174

1.144

Region(ref:Northeast)

Midwest

0.703**

0.706*

South

1.081

1.086

West

1.231

1.242

Con

stant

0.048***

0.047***

0.061***

0.062***

0.045***

0.044***

0.057***

0.059***

Observation

s128,921

128,921

128,921

128,921

128,921

128,921

128,921

128,921

Log

likelihood

&4.67

7e+0

6&4.676e+0

6&4.574e+0

6&4.552e+0

6&4.631e+0

6&4.62

7e+0

6&4.541e+0

6&4.519e+0

6

Source:Authors’calculations

ofMarch

Current

PopulationSurvey

(2008–

2013)and1980

decenn

ialCensus.Notes:***p

<0.001,

**p<0.01,*p

<0.05.


family than those in two-parent households with no significant differ-ences by socioeconomic status.

Living with Intermarried Parents

Intermarriage is a key indicator of assimilation. It is also an indicator thatby definition requires a sample restricted to households with two-parentmarried couples. Table 4 describes our sample of these households. It isnoteworthy that our sample is centered around children, so intermarriedcouples without children are excluded from the analyses.

Intermarriage is a more common parental circumstance for the thirdgeneration in 2010 relative to the second generation in 1980. While just 7.5

TABLE 4HOUSEHOLD CHARACTERISTICS FOR THE SECOND GENERATION IN 1980 AND FOR THE THIRD GENERATION

IN 2010 INCLUDE ONLY TWO-PARENT HOUSEHOLDS

Hispanic householdsAsian, non-Hispanic

householdsWhite, non-Hispanic

households

2ndGeneration(1980)

3rdGeneration(2010)

2ndGeneration(1980)

3rdGeneration(2010)

3rdGeneration(1980)

4th+Generation(2010)

Parents areintermarried (%)

7.5 28.5 20.0 62.0 3.1 9.3

Age (mean) 7 7 6 7 8 8Under age 10 (%) 69 67 78 70 55 58Head of householdage (mean)

37 36 38 39 37 39

Family size (mean) 5.5 4.8 4.7 4.5 4.7 4.6Number of children(mean)

3.2 2.6 2.4 2.4 2.7 2.5

Extended family (%) 16.9 8.1 20.6 9.6 4.0 3.5Fathers with BA+ (%) 7.7 21.8 51.4 56.3 24.7 42.3Mothers with BA+ (%) 4.6 24.8 40.7 57.7 14.0 44.8Unemployed headof household (%)

6.3 6.3 2.1 4.1 3.5 4.7

Working adults perhousehold (mean)

1.5 1.5 1.6 1.6 1.5 1.6

Household income(median, 2010 dollars,thousands)

$48 $58 $78 $100 $70 $87

Poverty status(% belowpoverty line)

20.8 13.9 6.0 5.4 6.2 5.0

Observations(unweighted)

49,404 4,890 17,130 1,532 1,616,394 66,707

Source: Authors’ calculations of March Current Population Survey (2008–2013) and 1980 decennial Census.Figures by generation status limited to two-parent households with children aged 16 and under. Figures are limitedto households where head of household is between 25 and 54 years old.


percent of second-generation Hispanic children in 1980 were growing up inhouseholds with parents of different ethnoracial origins, that figure is 28.5percent for the third generation in 2010. Asians also show significant inter-generational growth in the proportion growing up in households headed byan intermarried couple. In 1980, 20 percent of second-generation Asianchildren lived in such households, but in 2010, 62 percent of the third gen-eration living with married parents had parents of different ethnoracial ori-gins. Recall that we use a conservative estimate of intermarriage, countingonly as intermarriage unions between Hispanics and non-Hispanics; andbetween Asians and non-Asians. Thus, the significant rise in intermarriagebetween the two periods should be regarded as particularly significant.

The findings here should also be regarded as significant in view ofimmigration-driven compositional changes in population that might sig-nificantly limit intergenerational changes in the proportion of Asians andHispanics growing up in intermarried households. Immigration from Asiaand Latin America after 1980 increased dramatically, changing the oppor-tunity structure (Blau 1977) such that Asians and Hispanics have evengreater chances of findings marriage partners of the same origin. Otherresearch shows that large waves of Asian and Latin American immigrationdid indeed slow intermarriage rates for the second generation (Qian andLichter 2011). In spite of these immigration-driven compositionalchanges, which likely limited intermarriage, we still uncovered a dramaticincrease in the proportion of third-generation children, relative to the sec-ond generation 30 years earlier, that is living in households with intermar-ried parents in spite of ongoing immigration. While whites serve as thesocietal standard in these analyses, it is important to note that the propor-tion of children growing up in households with white parents who have anonwhite parent has tripled, from 3.1 percent of the third-plus generationin 1980, to 9.3 percent of the fourth-plus generation in 2010.

Using this same sample of two-parent households, we now turn tomultivariate analysis to examine the predictors of living with intermarriedparents. Table 5 includes the odds ratios of children in Hispanic house-holds (Models 1 and 2) and in Asian households (Models 3 and 4) livingwith intermarried parents in 1980 and 2010, relative to whites in thesame time periods.8

8Age is included in the model to control for differences in age distribution across groups.In Model 1, the age variable indicates that children at older ages are less likely to be inintermarried households than those at younger ages.


Model 1 shows that children in 2010 are 2.85 times more likely tolive in an intermarried household than those in 1980. Children with atleast one Hispanic parent are 2.46 times more likely to be living withintermarried parents than the white reference group. By 2010, third-gen-eration children have 1.6 times greater odds of living in householdsheaded by an intermarried couple relative to second-generation childrenin 1980. The results also show that children with college-educated moth-ers have 1.3 times the odds of living in an intermarried household relativeto those children living with mothers who have less than a bachelor’sdegree. Results show that geographic differences do not significantly shapethe odds of living in intermarried households for Hispanic generations rel-ative to whites. The only statistically significant geographic difference ischildren in the Midwest who have lower odds of living in intermarriedhouseholds compared to children in the Northeast.

When the analysis shifts to the sample that includes children livingwith at least one Asian parent, Model 3 shows that all children in 2010have roughly 3.59 greater odds living with intermarried parents than in1980. And the odds of living in an intermarried household are 8.1 timesgreater for children with at least one Asian parent than for whites. Similarto households with at least one Hispanic parent, the odds of living in an

TABLE 5ODDS RATIOS OF HAVING INTERMARRIED PARENTS (TWO-PARENT HOUSEHOLDS), CHILDREN WITH AT

LEAST ONE HISPANIC PARENT OR AT LEAST ONE ASIAN PARENT

Children withat least

1 HispanicParent

Children withat least 1HispanicParent

Children withat least 1 Asian

Parent

Children withat least 1 Asian

Parent(1) (2) (3) (4)

Year (1 = 2010) 2.852*** 2.834*** 3.592*** 3.484***Generation (1 = 3rd generation in2010; 2nd generation in 1980)

2.466*** 2.252*** 8.169*** 7.147***

Year 9 generation status 1.631*** 1.659*** 2.039*** 2.045***Age 0.974*** 0.975*** 0.981** 0.980**Mother’s education(1 = bachelor’s degree)

1.360*** 1.370*** 0.633*** 0.655***

Region (ref: Northeast)Midwest 0.808* 1.164South 0.932 1.508***West 1.198 2.076***

Constant 0.039*** 0.040*** 0.041*** 0.029***Observations 148,455 148,455 112,823 112,823Log likelihood &9.194e+06 &9.169e+06 &6.477e+06 &6.413e+06

Source: Authors’ calculations of March Current Population Survey (2008–2013) and 1980 decennial Census.Notes: ***p < 0.001, **p < 0.01, *p < 0.05.


intermarried household between second-generation children in 1980 andthird-generation children in 2010 increases among children with an Asianparent (2.03 times greater for the third generation). Unlike Hispanics,Model 3 shows that third-generation Asian children with a college-edu-cated mother have lower odds of living with intermarried parents thanthose with a mother who has less than a college education. Additionally,in contrast to Hispanics, there are significant geographic differences inintermarriage rates for those children living with at least one Asian parent.In Model 4, children living in the South and the West have greater oddsof living in an intermarried household than children in the Northeast.

Ethnoracial Identification of Children in Intermarried Households

When parents are intermarried, they may see themselves as having a widerarray of choices for how to label their children on surveys like the USCensus or CPS. We cannot infer too much from these labels about theidentity of the children, for their parents chose the categories for the chil-dren, and the category labels themselves may not capture the complexitiesof a lived ethnoracial identity (Dowling 2014). Nonetheless, the labels

TABLE 6RACIAL/ETHNIC IDENTIFICATION OF CHILDREN IN INTERMARRIED HOUSEHOLDS BY GENERATION

Child racial/Ethnic identification

One Hispanic parent Hispanic Non-Hispanic Share of Couples

1980Intermarried Hispanics 62.2% 37.8% 100%Mother, Hispanic 59.8% 40.2% 100% 61.6%Father, Hispanic 65.9% 34.1% 100% 38.4%

2010Intermarried Hispanics 75.6% 24.4% 100%Mother, Hispanic 74.2% 25.8% 100% 49.4%Father, Hispanic 76.9% 23.1% 100% 50.6%

One Asian parent Asiana Non-Asian Shared of couples

1980Intermarried Asians 44.4% 55.6% 100%Mother, Asian 44.6% 55.4% 100% 83.1%Father, Asian 43.7% 56.3% 100% 16.9%

2010Intermarried Asians 81.9% 18.1% 100%Mother, Asian 82.9% 17.1% 100% 68.1%Father, Asian 79.6% 20.4% 100% 31.9%

Source: Authors’ calculations of March Current Population Survey (2008–2013) and 1980 decennial Cen-sus.Notes: aIncluding those who were identified as Asian only and as Asian in combination with one or more other races.


that parents select for their children are suggestive of broader trends inthe salience of ethnoracial identities across populations. Using a samplerestricted to children living with intermarried parents, Table 6 reports eth-noracial labels by the combination of parents and the frequency of thelabels these parents give to their children. We show rates separately forwhich parent is of a particular ethnoracial background to determinewhether gender plays a role in determining the ethnoracial label given totheir children.

Among intermarried couples with a Hispanic parent in 1980, 62percent of second-generation children were given a Hispanic label. In2010, third-generation children living with one Hispanic and one non-Hispanic were labeled Hispanic more often than second-generation chil-dren three decades earlier. Among third-generation children in 2010, 75percent are given a Hispanic label. In both 1980 and 2010, althoughmore pronounced in 1980, those children with Hispanic fathers weremore likely to given a Hispanic label than those with a Hispanic mother.The increase in identifying their children as Hispanic could be associatedwith an increasing intermarriage rate for Hispanic men. However, there isnot a statistically significant difference in the use of the Hispanic labelwhen the mother is Hispanic versus when the father is Hispanic.

Asian children living with intermarried parents show different pat-terns in some respects. Among second-generation children in 1980, whenone parent was Asian and the other was non-Asian, children were givenan Asian label 44 percent of the time. Thus, second-generation childrenin 1980 growing up in intermarried households that had at least oneAsian parent were less often given an Asian label than children going upin Hispanic households with at least one Hispanic parent were given aHispanic label in 1980. Turning to 2010, it is important to recall thatparents in 2010 could use the CPS to indicate that their children hadmultiple racial backgrounds instead of choosing just one race category, asthey were forced to do in 1980. Thus, parents had the opportunity toidentify their children with both their Asian and non-Asian ancestries in2010. What we report from 2010 should be read as the proportion ofthird-generation children living in an intermarried household who weregiven any Asian (whether alone or in combination with another race cate-gory) label by their parents. Looking over time and across generations,patterns of identification of children living in Asian intermarried house-holds are similar to their Hispanic counterparts. Third-generation childrengrowing up in intermarried households with at least one Asian parent in


2010 were much more often given an Asian label than in 1980, and rela-tive to not being given an Asian label at all. Indeed, 81 percent of thesechildren were given an Asian label. It is important to emphasize theresults are almost surely driven in part by the option to choose more thanone category in 2010. Had that option been available in 1980, theincrease we observe might not appear so dramatic. Still, we observe a pat-tern of not just retention of an Asian or Hispanic label for children grow-ing up in intermarried households, but also a pattern of growingprevalence of those choices over time and across generations.

Turing to a multivariate analysis, we model the determinants of theethnoracial labels that parents assigned to their children. Table 7 displaysthe results as odds ratios.

Looking at the results for intermarried households with at least oneHispanic parent, third-generation children in 2010 have 1.5 times greaterodds of being labeled Hispanic relative to second-generation childrengrowing up in intermarried households in 1980. Older children see only aslight, but statistically significant reduction in their odds of being labeledHispanic. The gender–ethnoracial combination of parents has no

TABLE 7ODDS RATIOS PREDICTING CHILD’S ETHNORACIAL IDENTIFICATION AMONG CHILDREN IN INTERMARRIED

HOUSEHOLDS

CHILD identified as Hispanicamong intermarriedhouseholds with

one Hispanic parent (1)

CHILD identified as Asianamong intermarriedhouseholds with

one Asian parent (2)

Year (1 = 2010) 1.496*** 4.569***Age 0.979** 0.962*Mother’s race(ref: white, non-Hispanic)Same as child (Hispanic) 0.888Same as child (Asian) 1.436Other race 1.185 1.775

Mother’s education(1 = bachelor’s degree)

0.809* 1.335

Poverty status(1 = below poverty line)

0.908 0.496

Region (ref: Northeast)Midwest 0.980 2.590**South 0.828 1.772West 1.038 3.181***

Constant 2.845*** 0.283***Observations 38,440 8,598Log likelihood &3.563e+06 &1.010e+06

Source: Authors’ calculations of March Current Population Survey (2008–2013) and 1980 decennial Census.Notes: ***p < 0.001, **p < 0.01, *p < 0.05.


statistically significant effect. Mother’s education has only a slight effect,wherein having a mother with a bachelor’s degree reduces the odds ofbeing labeled Hispanic by roughly 20 percent. Poverty status does not havea statistically significant effect. Lastly, geographic differences are not signifi-cant.

A somewhat similar pattern emerges in intermarried households withat least one Asian parent (column 2). Third-generation children growingup in these households have odds of being labeled Asian, either alone orin combination with another racial category, that are 4.56 times greaterthan the odds for second-generation children in such households in1980.9 Intermarried couples with an Asian mother are not significantlymore likely to label their children as Asian than intermarried couple withan Asian father. Secondly, the socioeconomic status of the household,proxied by mother’s education and poverty status, does not have a statisti-cally significant impact on labeling their child as Asian. There are signifi-cant regional effects: Children living in the Midwest are 2.59 times morelikely than those in the Northeast of being labeled Asian. It is even higherfor those in the West (3.18 times greater odds) relative to the Northeast.

DISCUSSION

What is clear from these analyses is that the new third generation, com-pared to their second-generation parents 30 years earlier, is growing up inhouseholds with parents who have more education, are less often livingbelow the poverty line, earn more, are more often intermarried, and theyare less likely to live with extended family members. Those intergenera-tional differences suggest a pattern of assimilation that conforms to a clas-sic, straight-line model developed to explain the assimilation of Southernand Eastern European immigrants. Although the direction of the patternswe identify shows more similarities than differences between Hispanics

9It is important to note here that our supplementary analysis (available upon request)shows that the increase in the proportion of children who receive a nonwhite label acrossgeneration cohorts is not driven by an increase in marriages between nonwhites. Indeed,the overwhelming majority of intermarriage combinations in 1980 involve whites andAsians (28%) or whites and Hispanics (45%). Those percentages decline only slightly in2010 (38% and 23%, respectively). Marriage between Hispanics and people who identifiedthemselves as Other accounted for 11 percent of all intermarriages in 1980 and 18 percentin 2010. Marriage between Asians and people who identified as Other grew from 4 per-cent in 1980 to 6 percent in 2010.


and Asians, there are important group differences. In general, Asians in1980 had higher socioeconomic attainment than any group, includingnative-born whites, and they generally experienced an increase in theirsocioeconomic circumstances by 2010. Those differences are perhaps notsurprising given that the first-generation immigrants of these groups havevery different starting points. Indeed, large segments of the Asian immi-grant population are “hyperselected”: they have higher levels of educationand income than nonmigrants in their home country and are higher thanthe average individual in the United States (Lee and Zhou 2015). Theopposite is true for Mexicans, who make up a majority of Hispanics. Therelative position of the new third generation, as indicated by the kinds ofhouseholds they grow up in, reflects those starting points. The differencesalso show up in group disparities in the proportion living with two par-ents.10 Households with the Asian second generation in 1980 and thirdgeneration in 2010 are more likely to be headed by two parents thanhouseholds with Hispanic children and white, non-Hispanic children. Butgeneration-cohort changes within groups show slight declines over time.There is only a modest drop of two percentage points (or 75–73%) in theproportion of third-generation Hispanics living with both parents com-pared to second-generation Hispanics in 1980. The same is true forAsians — there is just a five percentage points (93–88.2%). It is notablethat these declines are also evident among native-born whites, who exhibitthe largest drop among the three groups at eight percentage points (88.0–80.9%) in the number of children living with both parents in 2010 com-pared to 1980. The decline in the share of two-parent households acrossall groups suggests a shift in societal norms around the timing of marriageand childbearing rather than a shift arising from socioeconomic or culturaldifferences. But overall, the household circumstances of the new thirdgeneration do not suggest a “third-generation decline” in which the for-tunes of the third generation fall behind the fortunes of the second gener-ation, as some have feared (see Huntington 2004). In fact, there is a clearpattern of improvement on virtually every measure in the household cir-cumstances of the new third generation as compared to the second genera-tion three decades prior. Of course, our results cannot speak to individualadult outcomes for the new third generation, which may yet evince a

10See McLanahan and Sandefur (1994) for an analysis of the effect of growing up with asingle parent.


third-generation decline. But if household circumstances are any indicator,such a decline does not appear to be on the horizon.

More opaque is the significance of changes in patterns of the identi-fication of children when they live in intermarried households. Recallfrom our findings that over time and across generations, third-generationchildren in 2010 who were growing up in a household with one Hispanicand one non-Hispanic parent had greater odds of their parents assigningthem a Hispanic label than second-generation children in intermarriedhouseholds three decades before; likewise third-generation children withone Asian and one non-Asian parent had greater odds of having their par-ents assign them an Asian label than a second generation in an intermar-ried household 30 years earlier. It is important to keep in mind thatparents in 2010 did not have to choose between an Asian label and anon-Asian label; they could choose multiple racial categories to identifytheir children. In fact, the most popular way to identify these childrenwas to select multiple racial categories, including one of the Asian sub-groups (60%). Existing literature suggests that those patterns may haveseveral, non-mutually exclusive explanations. It could be that the popular-ity of identifying children with multiple racial categories among thesehouseholds owes to growing recognition of mixed-race identity as a dis-tinctive form of ethnoracial identity. Interviews of individuals who haveone white and one Asian parent show that they express flexibility in thekinds of identities that they claim and express (Lee and Bean 2010).Among the flexible options is a distinctly multiracial identity that hasgained growing public recognition over the last two decades (DaCosta2007). Another possibility is that these patterns reflect the coming of asymbolic ethnicity — “a nostalgic allegiance to the culture of the immi-grant generation, or that of the old country; a love for and pride in a tra-dition that can be felt without having to be incorporated in everydaybehavior” (Gans 1979, 9) — among the post-1965 immigrants. Ampleresearch shows that an ethnoracial identity remained among the later gen-eration southern of the Eastern European immigration wave, but that eth-noracial identity was a dim part of individual and collective identity (Alba1990; Waters 1990). Among the new third generation, it could be thatclaimed ethnoracial attachments are symbolic and not deeply felt.

Finally, it may be that an institutionalized value of diversity and multi-culturalism encourages a strong attachment to an ethnoracial identity suchthat the new third generation will remain strongly tied to a nonwhite ethno-racial identity (Alba and Nee 2003; Bean, Brown, and Bachmeier 2015). As


the Southern and Eastern European-origin third generation was coming ofage, a nascent form of multiculturalism shaped a desire for them to be moreattached to an ethnoracial identity, even if that attachment ultimately turnedout to be symbolic (Jacobson 2006). Assimilation today may no longer becharacterized by a strong, positive correlation between intergenerationalchange in social, political, and economic dimensions of life. Postindustrialsocieties, like the United States, allow far greater latitude of individual iden-tities, including ethnoracial identities, and even design institutional policyto embrace, celebrate, and sometimes reward wide-ranging expressions ofthose identities (Bean, Brown, and Bachmeier 2015). A positive valuation ofethnoracial identification is embodied in an ideology of multiculturalismthat infuses new civic traditions of American identity premised on recogni-tion and pride in the varied immigrant origins that make up American iden-tity and culture (Schildkraut 2010).

On top of this ideology, patterns of contemporary immigration areat play. Ethnographic work suggests that the large-scale immigration fromLatin American and Asia gives the US-born individuals who share an eth-noracial origin with these immigrants greater access to ethnoracial symbolsand practices that make people feel more attached to an ethnoracial iden-tity. This “replenishment” (Jim!enez 2010) is especially true for the Mexi-can-origin population, which saw dramatic immigration-driven growthbetween the periods from which we drew our data. Indeed, it may be thatimmigration-driven compositional changes in the constituent groups thatmake up the Hispanic category are partly responsible for our findings. Insupplementary analyses not reported in the tables, we find that Mexicanchildren made up 67 percent of Hispanic children in second-generation-headed households with a Hispanic parent in 1980 and 74 percent ofHispanic children in third-generation-headed households with a Hispanicparent in 2010; while the Puerto Rican share fell from 11 percent to 5percent and the Cuban share also fell from 4 percent to 2 percent. OtherHispanics were 18 percent in 1980 and 19 percent in 2010. Likewise, itcould be that group-specific compositional changes among the largestAsian subgroups drive patterns among Asians. Unfortunately, the CPSdoes not identify individual Asian subgroups until 2013, the final year ofour CPS sample.11 The result of these ideological and immigration-driven

11These compositional changes in the Hispanic population are not a result of new immi-gration. Our analyses in 2010 include only the US-born, and so the compositional changesare due entirely to fertility and mortality.


demographic changes may be assimilation to an ethnoracially fleckedmainstream in which belonging is premised on recognition of and identi-fication with an immigrant ethnoracial ancestry (Alba and Nee 2003;Bean, Brown, and Bachmeier 2015; Jim!enez 2017).

These scenarios are speculative, of course. The census and CPS arenot well suited for inferring the meaning of ethnoracial identity of thenew third generation, as the data reflect choices parents made about howto identify their children and survey categories cannot capture the com-plexities of a lived identity. But the patterns that we observe suggest thatthe new third generation, in spite of having multiple ancestries, are notpulling away from an attachment to the ancestry of their immigrantgrandparents. Even if there is an alloyed version of that attachment, indi-cated by parents selecting multiple racial categories, an attachment enduresnonetheless.

A RESEARCH AGENDA FOR THE NEW THIRD GENERATION

This analysis of the new third generation offers a more temporally com-plete glimpse of post-1965 assimilation, but it is far from exhaustive. Ouranalysis ought to raise more questions than answers, spurring assimilationresearchers to understand the new third-generation outcomes in far greaterbreadth and depth. Our analyses, in combination with our reading of theexisting literature, suggest a research agenda for the next generation ofassimilation research. This agenda points toward the need for a significantdata-gathering effort as well as careful consideration of factors likely toaffect the third generation.

Tracking the New Third Generation

As our analysis indicates, there are severe data limitations in examiningthe new third-generation experience. Large government surveys such asthe decennial Census do not enable the identification of a third generationat all, asking only whether the individual filling out the Census and theother members of the household are US-born.12 The CPS data uponwhich we draw confines the identification of the new third generation to

12Although the American Community Survey calculates parent and child nativity, themicrodata include no information on whether a parent’s “own children” are US-born, norare these microdata publicly available.


children residing in the same household as their second-generation par-ents. Moreover, the CPS has a limited number of variables that scholarscan use to examine important assimilation outcomes, and the samplingstrategy makes examining important factors, such as those related to placeof residence, extremely difficult, if not impossible. There are other sourcesof data that capture a unique third and even fourth generation, such asthe Mexican American Study Project (Ortiz and Telles 2011) and theImmigration and Intergenerational Mobility in Metropolitan Los Angeles(Rumbaut et al. 2008). But because these two data sets include adults,they capture the third- and later-generation descendants of a previouswave of Mexican immigrants, and not of the post-1965 immigrationwave.

What is required to advance an understanding of the new third gen-eration, then, is data that allow for the identification of the post-1965third generation as adults and that include information on the kinds ofindividual-level assimilation outcomes that other standard surveys lack.Existing data-gathering efforts on the second generation — the Childrenof Immigrants Longitudinal Survey (CILS) and the Immigrant SecondGeneration in Metropolitan New York (ISGMNY) — are models for sim-ilar endeavors that could be emulated to study the new third generation.CILS samples more than 5,000 children of immigrants in the Miami andSan Diego metropolitan areas in 1992, when these children were in theirearly teens, and re-interviewed most of the original sample three yearslater, and again in 2001–2003. CILS also includes parent surveys and in-depth interviews with select survey respondents. ISGMNY is a survey andinterview sample of 1.5 (arrived before the age of 12)- and second-genera-tion individuals living in the New York metropolitan area. The samplealso includes native-born comparisons groups, including whites, PuertoRicans, and African Americans.

The features of these studies that distinguish each from the other —longitudinal design; non-post-1965 comparison groups — as well as simi-larities across the studies — qualitative and quantitative data; contextualdata; a range of assimilation measures — could be incorporated into asingle study of the new third generation.

The New Geography of Assimilation

There are also more substantive imperatives for studying the new thirdgeneration. A notable shortcoming of existing second-generation studies is


a decided lack of attention to the second-generation experiences in newimmigrant gateways in the US South and Midwest. Although there hasbeen a spate of rich studies examining the dynamics of settlement andintergroup relations in the new gateways, notably absent are studies thatexamine the second-generation experience in these new gateways (for anotable exception, see Hern!andez-Le!on and Lakhani 2013; Tran and Val-dez 2015). That empirical omission is even more glaring with the rise ofthe third generation. Assimilation scholars have pointed out that there isgood reason to believe that the immigrant and second-generation experi-ences may differ in the new gateways because of the unique economic,political, and social histories of these regions (Waters and Jim!enez 2005;Tran and Valdez 2015; Marrow 2009; Massey 2008). Indeed, the geogra-phy variables show up as significant predictors in each of our multivariatemodels. Although these variables are blunt instruments for gaining lever-age on the effect of geographic context, they do show that assimilationoutcomes are unfolding at different speeds, depending on where the newthird generation is growing up. Explaining exactly how the regional differ-ences are reflected in intergenerational assimilation into the third genera-tion remains a pressing area of inquiry, especially given the regionalizationof the economy and the related educational and economic opportunities(Chetty et al. 2014b).

The New Third Generation as a Cohort

The new third generation will be coming of age under historical circum-stances that may shape their experiences in very particular ways. Thesecohort-specific conditions are especially meaningful when making compar-isons between third-generation individuals descended from previous wavesof immigrants. For example, the third-generation descendants of theSouthern and Eastern European immigrant wave were mostly baby boom-ers who came of age at a time of relative economic equality (compared toother points in US history, including today), massive social upheaval, andvery little new immigration. Those historical events almost certainlyformed what it meant to be a grandchild of immigrants in ways thatshaped assimilation outcomes.

Likewise, today’s new third generation will come of age as a groupof individuals who have in common not just the experience of being thirdgeneration but also being part of a cohort that experiences a set of histori-cal events at similar stages in the life course. The specifics of those events


have yet to unfold, but there are social and economic trends underwaythat are likely to shape the extent and kind of their assimilation: highincome and wealth inequality; contestation over multiculturalism anddiversity policies; a two-term black president; the Trump presidency; massimprisonment; the Internet age; more liberal attitudes about interethnora-cial dating and marriage; multi-ethnoracial identity; low levels of militaryservice; disinvestment from public education; and the post-9/11 focus onsecurity. The effects of inequality may already show up in shifting normsabout age of marriage and childbearing, which, even in our data, shapethe household circumstances of Hispanics, Asians, and whites, althoughnot to equal degrees. The recognition and support of wide-ranging indi-vidual identities, including ethnoracial identities (Bean, Brown, and Bach-meier 2015), may also show up in our analysis in the greater propensityfor new third-generation children living with intermarried parents toreceive a nonwhite label in 2010, relative to a second generation livingwith intermarriage parents in 1980.

Gendered Paths of Assimilation

The overwhelming focus of assimilation research is on how ethnoracial ori-gin and generation structure assimilation outcomes. But previous researchon the second generation highlights how gender significantly shapes assimi-lation outcomes (Park and Myers 2010; Park, Nawyn, and Benetsky 2015).A pressing question for the next generation of assimilation research iswhether and how gender continues to shape socioeconomic outcomes aspart of the assimilation process. The new third generation may very well beswept up in more secular gendered patterns. But there may also be gen-dered dimensions of their lives related specifically to their third-generationstatus and ethnoracial origins. Our analysis points to two such dimensions.Among the households headed by a single parent, women head the major-ity. Hispanics far more often live in households headed by a single mothercompared to whites and Asians. And in households headed by two parentsof different ethnoracial origins, where one spouse is Asian, it is far moreoften the mother who is Asian. What, then, does it mean for group out-comes when so many third-generation Hispanics are being raised by a sin-gle mother? And how is the collective “mixed” childhood experience shapedby the fact that it is overwhelmingly the mother who is nonwhite and thefather is white? As our analysis and other work show, ethnoracial origin is


not the only category potentially marking differences in assimilation; gen-der may also be central to that segmentation.

The New Third Generation Followed by a New ImmigrantGeneration?

The new third-generation experience will also be determined by factorsendogenous to the ethnoracial groups that shape group composition. Thethird generation represents a temporal march away from the immigrantgeneration. But the new third generation may nonetheless continue to nav-igate contexts in which they have varying degrees of contact with immi-grants of the same ethnoracial origin. With few exceptions (see, e.g.,Erdmans 1998), Southern and Eastern European immigration all butended with the passage of the 1924 Immigration Quotas Act, whichseverely limited legal immigration from most parts of Europe. The decid-edly US-born heavy composition of Southern and Eastern European eth-noracial groups was a likely contributor to rising intermarriage ratesbecause there were simply fewer individuals of the same ethnoracial originto marry (Blau 1977), and it almost certainly shaped the symbolic formand content of ethnoracial identity (Waters 1990; Jim!enez 2010). Whetherimmigration from their grandparents’ homeland continues or abates willshape the exposure that the new third generation has to ethnoraciallylinked practices and symbols; the content of ethnoracial stereotypes; ethno-racial prejudice; and the composition of networks and the availability ofromantic partners (Tuan 1998; Jim!enez 2010; Tsuda 2016).

Our findings point toward one potential effect of these immigrationpatterns. Third-generation children growing up with intermarried parentshave greater odds of being assigned to the ethnoracial ancestry of theirimmigrant grandparents than their second-generation counterparts 30 yearsearlier. The questions remain: How does the new third generation constructand assign meaning to their ethnoracial identities? An important factor iswhether immigration continues, and whether the corresponding salience ofsome ethnoracial origins makes particular forms of identity more accessible.For the Mexican-American new third generation, the largest subgroup inthe new third generation, the future flows of Mexican immigration are verymuch in doubt. There is now negative net Mexican migration to the Uni-ted States (Gonzalez-Barrera 2015). In contrast, Asian immigration is onthe rise, and there are now more immigrants entering the United States


from Asia than from Latin America (Pew Research Center 2015). However,migration patterns unfold, those patterns will significantly shape how thenew third generation forms their ethnoracial identities.

CONCLUSION

Now is the time for assimilation scholars to turn their attention to thenew third generation — the grandchildren of the post-1965 immigrants— with the same vigor that these scholars have studied the new secondgeneration for the past two decades. Debates about the form and contentof the assimilation of the post-1965 immigrants are largely at a stalematebecause the focus on the second generation offers a temporally truncatedpicture of assimilation. With the rise of the new third generation, socialscientists have an unprecedented opportunity to examine, in detail, howthe next chapter of American assimilation is writing itself. We offer a firstlook. Our analysis suggests that assimilationist proceeding apace, with thenew third generation in 2010 living in households that are socioeconomi-cally better off, more likely to be headed by intermarried parents, and lesslikely to have extended family members present. We also find that identi-fication with the ethnoracial background of the immigrant grandparentsappears to endure, and may even be more prevalent in the third genera-tion. While instructive, our analysis hardly offered definitive evidence of anew third generation that may be on a path of assimilation similar to onethat previous waves traveled, or one that looks more like the segmentedassimilation paths identified in some studies of the second generation. Set-tling that debate requires the emergence of a new generation of assimila-tion research that further explores the questions implied by our analysisand offered in our research agenda.

REFERENCES

Alba, Richard D.1985 Italian Americans: Into the Twilight of Ethnicity. Englewood Cliffs, NJ: Prentice-

Hall.———1990 Ethnic Identity: The Transformation of White America. New Haven, CT: Yale

University Press.Alba, Richard, Philip Kasinitz, and Mary C. Waters2011 “The Kids are (Mostly) Alright: Second-Generation Assimilation: Comments on

Haller, Portes and Lynch.” Social Forces 89:763–73.


———, and Victor Nee2003 Remaking the American Mainstream: Assimilation and Contemporary Immigration.

Cambridge, MA: Harvard University Press.Bean, Frank D., Susan K. Brown, and James D. Bachmeier2015 Parents Without Papers: The Progress and Pitfalls of Mexican American Integration.

New York, NY: Russell Sage Foundation.Blau, Peter M.1977 Inequality and Heterogeneity: A Primitive Theory of Social Structure. New York, NY:

Free Press.Chetty, Raj, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez2014a “Where Is the Land of Opportunity? The Geography of Intergenerational Mobility

in the United States.” Quarterly Journal of Economics 129 (4):1553–623.———, ———, ———, ———, and Nicholas Turner2014b “Is the United States Still a Land of Opportunity? Recent Trends in Intergenera-

tional Mobility.” American Economic Review 104 (5):141–47.DaCosta, Kimberly Mc Clain2007 Making Multiracials: State, Family, and Market in the Redrawing of the Color Line.

Stanford, CA: Stanford University Press.Dowling, Julie A.2014 Mexican Americans and the Question of Race. Austin, TX: University of Texas Press.Duncan, Otis D.1966 “Methodological Issues in the Analysis of Social Mobility.” In Social Structure and

Mobility in Economic Development, edited by Neil J. Smelser, and Seymour M. Lip-set, 51–97. Chicago, IL: Aldine.

Duncan, Brian, and Stephen J. Trejo2007 “Ethnic Identification, Intermarriage, and Unmeasured Progress by Mexican Ameri-

cans.” In Mexican Immigration to the United States, edited by George J. Borjas,229–68. Chicago, IL: University of Chicago Press.

———, and ———2011 “Tracking Intergenerational Progress for Immigrant Groups: The Problem of Eth-

nic Attrition.” The American Economic Review 101 (3):603–608.Emeka, Amon, and Jody Agius Vallejo2011 “Non-Hispanics With Latin American Ancestry: Assimilation, Race, and Identity

Among Latin American Descendants in the US.” Social Science Research 40(6):1547–63.

Erdmans, Mary P.1998 Opposite Poles: Immigrants and Ethnics in Polish Chicago, 1976–1990. University

Park, PA: Pennsylvania State University Press.Foner, Nancy2005 In a New Land: A Comparative View of Immigration. New York, NY: New York

University Press.Fox, Cybelle, and Thomas A. Guglielmo2012 “Defining America’s Racial Boundaries: Blacks, Mexicans, and European Immi-

grants, 1890–1945.” American Journal of Sociology 118 (2):327–79.Gans, Herbert J.1979 “Symbolic Ethnicity: The Future of Ethnic Groups and Cultures in America.” Eth-

nic and Racial Studies 2 (1):1–20.———1992 “Second-Generation Decline: Scenarios for the Economic and Ethnic Futures of

the Post-1965 American Immigrants.” Ethnic and Racial Studies 15(2):173–92.


Garcia, Ginny2011 Mexican American and Immigrant Poverty in the United States. Springer Series on

Demographic Methods and Population Analysis, vol. 28. New York, NY: SpringerScience & Business Media.

Gonzalez-Barrera, Ana2015 More Mexicans Leaving Than Coming to the U.S. Net Loss of 140,000 From 2009 to

2014; Family Reunification Top Reason for Return. Washington, DC: Pew ResearchCenter.

Gordon, Milton M.1964 Assimilation in American Life: The Role of Race, Religion, and National Origins. New

York, NY: Oxford University Press.Haller, William, Alejandro Portes, and Scott M. Lynch2011 “Dreams Fulfilled, Dreams Shattered: Determinants of Segmented Assimilation in

the Second Generation.” Social Forces 89 (3):733–62.Hern!andez-Le!on, Rub!en, and Sarah M. Lakhani2013 “Gender, Bilingualism, and the Early Occupational Careers of Second-Generation

Mexicans in the South.” Social Forces 92 (1):59–80.Huntington, Samuel P.2004 Who Are We? The Challenges to America’s National Identity. New York: Simon &

Schuster.Jacobson, Matthew F.2006 Roots Too: White Ethnic Revival in Post-Civil Rights America. Cambridge, MA: Har-

vard University Press.Jim!enez, Tom!as R.2010 Replenished Ethnicity: Mexican Americans, Immigration, and Identity. Berkeley, CA:

University of California Press.———2017 The Other Side of Assimilation: How Immigrants are Changing American Life. Oak-

land, CA: University of California Press.Kao, Grace, and Marta Tienda1995 “Optimism and Achievement: The Educational Performance of Immigrant Youth.”

Social Science Quarterly 76 (1):1–19.Kasinitz, Phillip, John H. Mollenkopf, Mary C. Waters, and Jennifer Holdaway2008 Inheriting the City: The Second Generation Comes of Age. Cambridge, MA & New

York, NY: Harvard University Press & Russell Sage Foundation.King, Miriam, Steven Ruggles, J. Trent Alexander, Sarah Flood, Katie Genadek, MatthewB. Schroeder, Brandon Trampe, and Rebecca Vick2010 Integrated Public Use Microdata Series, Current Population Survey: Version 3.0.

[Machine-Readable Database]. Minneapolis, MN: University of Minnesota.LaLonde, Robert J., and Robert H. Topel1992 “The Assimilation of Immigrants in the US Labor Market.” In Immigration and the

Workforce: Economic Consequences for the United States and Source Areas, edited byGeorge J. Borjas, and Richard B. Freeman, 67–92. Chicago, IL: University of Chi-cago Press.

Lee, Jennifer, and Frank D. Bean2010 The Diversity Paradox: Immigration and the Color Line in Twenty-First Century

America. New York, NY: Russell Sage Foundation.———, and Min Zhou2015 The Asian American Achievement Paradox. New York, NY: Russell Sage Founda-

tion.


Lichter, Daniel T., Julie H. Carmalt, and Zhenchao Qian2011 “Immigration and Intermarriage Among Hispanics: Crossing Racial and Genera-

tional Boundaries.” Sociological Forum 26 (2):241–64.Lieberson, Stanley, and Mary C. Waters1988 From Many Strands: Ethnic and Racial Groups in Contemporary America. New York,

NY: Russell Sage Foundation.Macias, Thomas2006 Mestizo in America: Generations of Mexican Ethnicity in the Suburban Southwest.

Tucson, AZ: University of Arizona Press.Marrow, Helen B.2009 “New Immigrant Destinations and the American Colour Line.” Ethnic and Racial

Studies 32 (6):1037–57.Massey, Douglas S.2008 New Faces in New Places: The Changing Geography of American Immigration. New

York, NY: Russell Sage Foundation.McLanahan, Sara, and Gary Sandefur1994 Growing Up With a Single Parent. What Hurts, What Helps. Cambridge, MA: Har-

vard University Press.National Academies of Science, Engineering, and Medicine2015 The Integration of Immigrants Into American, Washington, DC: National Academies

Press.Ortiz, Vilma, and Edward E. Telles2011 The Mexican American Study Project II (MASP II), 1998–2000. ICPSR28481-v2.

Ann Arbor, MI: Inter-university Consortium for Political and Social Research [dis-tributor], 2011-03-29. https://doi.org/10.3886/icpsr28481.v2

Park, Robert E., and Ernest W. Burgess1969[1921]

Introduction to the Science of Sociology. Chicago, IL: University of Chicago Press.

Park, Julie, and Dowell Myers2010 “Intergenerational Mobility in the Post-1965 Immigration Era: Estimates by an

Immigrant Generation Cohort Method.” Demography 47 (2):369–92.———, ———, and Tom!as R. Jim!enez2014 “Intergenerational Mobility of the Mexican-Origin Population in California and

Texas Relative to a Changing Regional Mainstream.” International MigrationReview 48 (2):442–81.

———, Stephanie J. Nawyn, and Megan J. Benetsky2015 “Feminized Intergenerational Mobility Without Assimilation? Post-1965 US Immi-

grants and the Gender Revolution.” Demography 52 (5):1601–26.Parrado, Emilio A., and S. Philip Morgan2008 “Intergenerational Fertility Among Hispanic Women: New Evidence of Immigrant

Assimilation.” Demography 45 (3):651–71.Perlmann, Joel2005 Italians Then, Mexicans Now: Immigrant Origins and Second-Generation Progress,

1890–2000. New York, NY: Russell Sage Foundation.———, and Roger Waldinger1997 “Second Generation Decline? Children of Immigrants, Past and Present - A Recon-

sideration.” International Migration Review 31 (4):893–922.


Pew Research Center2015 Modern Immigration Wave Brings 59 Million to U.S., Driving Population Growth

and Change Through 2065: Views of Immigration’s Impact on U.S. Society Mixed.Washington, DC: Pew Research Center.

Portes, Alejandro, and Rub!en G. Rumbaut2001 Legacies: The Story of the Immigrant Second Generation. Berkeley, CA and New

York, NY: University of California Press; Russell Sage Foundation.———, and ———2014 Immigrant America: A Portrait. Oakland, CA: University of California Press.———, and Min Zhou1993 “The New Second Generation: Segmented Assimilation and its Variants.” Annals of

the American Academy of Political and Social Science 530 (1):74–96.Qian, Zhenchao, and Daniel T. Lichter2007 “Social Boundaries and Marital Assimilation: Interpreting Trends in Racial and

Ethnic Intermarriage.” American Sociological Review 72 (1):68–94.———, and ———2011 “Changing Patterns of Interracial Marriage in a Multiracial Society.” Journal of

Marriage and Family 73 (5):1065–84.Ramakrishnan, Karthick S.2004 “Second-Generation Immigrants? The “2.5 Generation” in the United States.”

Social Science Quarterly 85(2):380–99.Ruggles, Steven, Katie Genadek, Ronald Goeken, Josiah Grover, and Matthew Sobek2015 Integrated Public Use Microdata Series: Version 6.0 [Machine-Readable Database].

Minneapolis, MN: University of Minnesota.Rumbaut, Rub!en G.2005 “Turning Points in the Transition to Adulthood: Determinants of Educational

Attainment, Incarceration, and Early Childbearing Among Children of Immi-grants.” Ethnic and Racial Studies 28 (6):1041–86.

———, Frank D. Bean, Leo R. Ch!avez, Jennifer Lee, Susan K. Brown, Louis DeSipio,and Min Zhou2008 Immigration and Intergenerational Mobility in Metropolitan Los Angeles (IIMMLA),

2004 ICPSR22627-v1. Ann Arbor, MI: Inter-University Consortium for Political andSocial Research [distributor], 2008-07-01. https://doi.org/10.3886/icpsr22627.v1

Schildkraut, Deborah2010 Americanism in the Twenty-First Century: Public Opinion in the Age of Immigration.

New York, NY: Cambridge University Press.Smith, James P.2003 “Assimilation Across the Latino Generations.” American Economic Review 93

(2):315–19.———2006a “Immigrants and the Labor Market.” Journal of Labor Economics 24 (2):203–33.Smith, Robert C.2006b Mexican New York: Transnational Lives of New Immigrants. Berkeley, CA: Univer-

sity of California Press.Telles, Edward E., and Vilma Ortiz2008 Generations of Exclusion: Mexican Americans, Assimilation, and Race. New York, NY:

Russell Sage Foundation.———, and Christina A. Sue2009 “Race Mixture: Boundary Crossing in Comparative Perspective.” Annual Review of

Sociology 35:129–46.


Tran, Van C., and Nicol M. Valdez2015 “Second-Generation Decline or Advantage? Latino Assimilation in the Aftermath of

the Great Recession.” International Migration Review 51 (1):155–90.Tsuda, Takeyuki2016 Japanese American Ethnicity: In Search of Heritage and Homeland Across Generations.

New York, NY: New York University Press.Tuan, Mia1998 Forever Foreigners or Honorary Whites? The Asian Ethnic Experience Today. New

Brunswick, NJ: Rutgers University Press.Warner, W. L., and Leo Srole1945 The Social Systems of American Ethnic Groups. New Haven, CT: Yale University

Press.Waters, Mary C.1990 Ethnic Options: Choosing Identities in America. Berkeley: University of California

Press.———2010 “Second Generation Advantages: Recasting the Debate.” In Helping Young Refugees

and Immigrants Succeed: Public Policy, Aid, and Education, edited by Gerhard Son-nert, and Gerald Holton, 89–100. New York, NY: Palgrave MacMillan.

———, and Tom!as R. Jim!enez2005 “Assessing Immigrant Assimilation: New Empirical and Theoretical Challenges.”

Annual Review of Sociology 31:105–25.———, Van C. Tran, Philip Kasinitz, and John H. Mollenkopf2010 “Segmented Assimilation Revisited: Types of Acculturation and Socioeconomic

Mobility in Young Adulthood.” Ethnic and Racial Studies 33 (7):1168–93.Zhou, Min, and Yang Sao Xiong2005 “The Multifaceted American Experiences of the Children of Asian Immigrants: Les-

sons for Segmented Assimilation.” Ethnic and Racial Studies 28 (6):1119–52.


the new third generation: post-1965 immigration and the next … · 2017-12-19 · assimilation...

Documents